Cognigy

Cognigy.AI Docs

COGNIGY.AI is the Conversational AI Platform focused on the needs of large enterprises to develop, deploy and run Conversational AI’s on any conversational channel.

Given the arising need of voice interfaces as the most natural way of communicating with brands, Cognigy was founded in 2016 by Sascha Poggemann and Phil Heltewig. Our mission: to enable all devices and applications to intelligently communicate with their users via naturally spoken or written dialogue.

Get Started

Annotations

The Cognigy.AI intent builder features the Annotations tool that allows sentence words to be assigned as placeholders to recognize specific details from user utterances. Each placeholder is assigned to the Slot that will be recognized and filled by the NLU.

This feature is particularly useful to reduce repetition by allowing lexicon libraries to be recognized within example sentence structures.

Adding Annotations

Annotations can be added to example sentences by highlighting the required word in the text field and selecting the "+" icon at the right end of the field.

A new menu line will be created for each annotation added to the sentence. The annotation menu includes drop down fields to select the Slot Type and key. Both of these values are required for each annotation.

Multiple annotations of different types can be added to a single example sentence where each annotation will be underlined in a different color.

📘

Copy and Paste Annotated Sentences

It is possible to copy a sentence including annotations to another example sentence line which will carry all of the assigned annotations with it. Although similar sentences should be avoided, this can help save time by reducing the need to re-annotate words.

Lexicon Slot

Insert Lexicon Slot placeholders within example sentences to teach the NLU to recognize words with similar meanings. For example, annotate the product name shirts with a custom product lexicon tag to allow the NLU to recognize pants, hats, and jackets that have been given the same slot tag in the Lexicon Editor.

Synonyms you provide in attached Lexicons are used at score and train time to improve accuracy.

Note that Slot Tags and Synonyms for intents must be attached in the same flow. This means when training with attached flows: intents in attached flows use the lexicons of their attached flows, not the lexicons of the parent Flow.

📘

Pro Tip: use synonyms to translate rare words

Provide understandable, common synonyms for rare words. Cognigy's Natural Language Understanding will be able to better make sense of rare words if you translate them to common utterances. E.g., a specific brand name such as "PremiumPayID" could be provided with the synonym "credit card" and Cognigy will understand its connection to payments where it would otherwise be an unknown word.

🚧

Do not overuse tags

For small word groups which mean the same thing it more effective to use only synonyms. Tags are appropriate for a large number of words collected in one concept, such as first names or country codes.

📘

Exact tag replaced matching algorithm

Example and input utterances that match exactly save for Keyphrase(s) which share the same Tag automatically receive a score of 0.95

To illustrate:

The input "I want a pizza" will match an example sentence "I want a burger" directly with score 0.95 if "pizza" and "burger" are Keyphrases in an attached Lexicon that share the same Tag.

Note the direct comparison is limited to the very first Tag in a Lexicon only for efficiency reasons.

System Slot

Cognigy.AI features built-in slot detection for a broad range of slots. The system slots that are available to add in the annotations drop down menu are:

Slot Name

Number

Date

Age

Temperature

Duration

Percentace

E-Mail

URL

Money

Distance

📘

Detailed Slot Information

For more information about System Slots, including data formats and examples, visit the Slot Mapping.

Any Slot

This slot option provides an open text field for a custom slot name to be specified. The name should match any slot that will be recognised by the NLU and published to the input data as a child of input.slots.

Updated about a year ago


What's Next

Intent Analyzer

Annotations


Suggested Edits are limited on API Reference Pages

You can only suggest edits to Markdown body content, but not to the API spec.